PR Review Squad
by NeuroSquad teamVersion 1updated Oct 2, 2026
The app shows everything it will install and asks once before anything runs.
The canvas
About this template
Three AI reviewers from three different model families look at the same change, and a lead merges what they find into one review you can act on.
What's inside
- Lead reviewer (Claude Code) reads the diff, hands the same scope to the other two reviewers over its arrows (
agent_send/agent_wait), double-checks every finding and writes the final review. - Second reviewer (Codex CLI) and Third reviewer (Gemini CLI) give independent opinions: correctness and edge cases; tests, security and readability.
- Code Graph builds a knowledge graph of the repository (functions, calls, imports) so all three can find callers and dependents of changed code without grep loops. Download its engine (about 40 MB, checked by SHA-256) from the card with one click.
- Review notes is the note the lead writes into, with a ready skeleton: Blocking, Should fix, Nits.
- GitHub (optional) is GitHub's official MCP server (hosted), for PR comments and metadata. It needs a token: type
Bearerand a personal access token when asked, or leave it empty and ignore that card.
How to use
- Import into the repository you want reviewed and check out the branch.
- The lead's first prompt is pasted, not sent: change the scope (a branch, a PR number, a folder) and press Enter.
- The reviewers are told not to modify files; the result lands in Review notes.
Each agent runs on its own CLI login. CLIs you don't have yet are offered for install in the import dialog; you can also delete a reviewer card and keep the rest.
More templates
See all- OfficialFeaturedCoding
Long-Running Project with Memory
by NeuroSquad team2 agents · 5 cards
For a project that lasts weeks: decisions, conventions and gotchas go into a shared memory, and come back automatically in later sessions and to the other agent. What's inside…
- #memory
- #mem0
- #long-running
- #multi-agent
1 like1 download - OfficialFeaturedCoding
Bug Hunt Loop
by NeuroSquad team1 agent · 5 cards
One agent that reproduces a bug before it touches the code, finds the root cause, proves the fix with a test and writes down what happened. What's inside Fixer (Claude Code) with…
- #debugging
- #bug-fix
- #testing
- #browser
- OfficialFeaturedWriting
Docs Writer with Live Library Docs
by NeuroSquad team1 agent · 4 cards
A documentation writer that checks every library it mentions against current docs instead of its training data, so examples use APIs that actually exist in your versions. What's…
- #documentation
- #context7
- #writing
Put your squad to work
Free for Windows and macOS. Bring the agents you already use.
Linux is coming later.
The first-run wizard can install an agent CLI, Git and the dictation model for you — or skip each step.